Bulk and single-cell transcriptomic profiling of lipid metabolism identifies ARSB as a potential target for suppressing colorectal cancer growth and immune evasion

Tumors frequently remodel their metabolic programs to support proliferation while suppressing cell death and anti-tumor immune responses. However, the regulatory mechanisms linking metabolic rewiring to immune evasion remain incompletely understood. In this study, we developed an integrative machine learning framework to construct a lipid metabolism-related signature (LMRS). The LMRS was an independent risk factor for overall survival (OS) and showed stable predictive performance across validation cohorts. An LMRS-based nomogram further provided a quantitative tool for individualized prognostic assessment. High- and low-risk groups differed in biological pathway activity, mutational landscape, and immune cell infiltration within the tumor microenvironment (TME). Among the LMRS genes, ARSB emerged as a candidate tumor suppressor in colorectal cancer (CRC). ARSB protein was downregulated in CRC and inhibited tumor growth, at least partly by attenuating ERK signaling. ARSB also enhanced MHC-I-associated antigen-presentation programs and reduced PD-L1 expression, thereby increasing CD8 + T cell-mediated cytotoxicity. In vivo, ARSB overexpression increased CD8 + T cell infiltration and activation. These findings suggest that ARSB constrains CRC progression through coordinated regulation of ERK signaling and CD8 + T cell activity.